College admission tests aim to predict students’ future academic performance by assessing relevant cognitive abilities and competencies. Such tests are frequently used for student selection purposes in Germany, Austria, and Switzerland. Still, test contents, test takers, legal aspects and the educational system are not static, making test development and use an ongoing challenge. A first meta-analysis by Hell et al. (2007) found a substantial predictive validity for subject-specific tests in German-speaking countries with the highest validities for the field of human medicine. The present study aims to replicate and extend these previous findings using newer publications from 2005 to 2018 (K = 54 independent samples, N = 8,410 persons). The new meta-analysis finds a correlation of ρ = .43 (corrected for range restriction and for criterion reliability) between test scores and subsequent academic grades, confirming the predictive power of subject-specific college admission tests. Previous findings are replicated and extended to other fields. Good predictive validities are found in the fields of linguistics, law, social sciences, medical studies, engineering, natural sciences, economics, and information sciences. Moderator analyses show that higher validities were associated with newer studies, articles with peer-review, university student samples, and a shorter time span between the test and the assessment of college grades. Compared to international studies, the tools used at German-speaking universities appear to have a similar or even higher predictive validity.